An article in American Scientist credits the past work of SFI External Professor Joseph Traub and his former PhD student Spassimir Paskov with bringing about a "dramatic revival of interest" in the quasi-Monte Carlo method of deterministic sampling, useful for computing 360 dimensional integrals such as those occurring in financial computations.
Traub, the Edwin Howard Armstrong Professor of Computer Science at Columbia University, and Paskov in the early 1990s showed through computer experimentation that quasi-Monte Carlo, which involves deterministic sampling, beats Monte Carlo, which involves randomized sampling, by one to three orders of magnitude for computing the 360 dimensional integrals, which occur in computational finance. The use of quasi-Monte Carlo for high dimensional integrals was counter to the conventional wisdom of the world's leading experts.
Theoretical explanations of these results continue to be an active research area, says Traub. There is no generally accepted explanation.
Read the American Scientist article (July-August 2011)
More about Traub's quasi-Monte Carlo method